Factors Associated With Early Discontinuation of Radiation Therapy: An Analysis of the National Cancer Database

Abstract

<h3>Purpose/Objective(s)</h3> Radiation courses generally involve multiple visits over the course of weeks, depending on the disease site and treatment technique. Therefore, radiation therapy (RT) may be discontinued prior to planned treatment completion. The purpose of this analysis is to identify any clinical or socioeconomic factors that could serve as predictors of RT discontinuation. <h3>Materials/Methods</h3> Using the National Cancer Database (NCDB) data from 2017, we identified 7,561 cases treated with RT, chemoradiation (CRT), surgery with RT (S+RT), or surgery with CRT (S+CRT) that had information on radiation discontinuation. The variables that we assessed include age (18- < 50, 50- < 70, 70+), gender (male, female), race (White, Black, other), insurance status (private, Medicare, Medicaid/uninsured, other government), income level (< $30,000, $30,000- < $35,000, $35,000- < $46,000, $46,000+), Charlson-Deyo Comorbidity Score (0, 1, ≥2), treatment type (RT, CRT, S+RT or S+CRT), disease site, and distance from patient residence to treatment facility. RT discontinuation was compared to socioeconomic and clinicopathologic characteristics using chi-square analysis for categorical predictors and Wilcoxon-Mann-Whitney test for continuous predictors. <h3>Results</h3> Of the 7,561 patients with information on RT discontinuation, 7,349 (97.2%) patients completed RT, whereas RT was discontinued in 212 (2.8%) patients. The reasons for discontinuation are shown in Table 1. Patients who are 70+ years old (3.7%, <i>P</i> = 0.008), male (3.4%, <i>P</i> = 0.014), Black (3.2%, <i>P</i> = 0.030), Medicaid/uninsured (3.9%, <i>P</i> < 0.0001), income < $30,000 (4.9%, <i>P</i> < 0.0001) and Charlson-Deyo Comorbidity Score ≥2 (3.9%, <i>P</i> = 0.005) had the highest percentages of RT discontinuation in their respective categories. By treatment type, CRT (5.94%) had the highest discontinuation rates, followed by RT (3.1%), S+CRT (2.1%), and S+RT (1.8%), <i>P</i> < 0.0001. Disease site correlated with RT discontinuation (<i>P</i> < 0.0001) with the highest rates of RT discontinuation for cancers of the urinary tract (8.8%), head and neck (6.2%), and respiratory tract (6.0%). The median distance between patient residence and treatment facility was 11.2 and 9.6 miles for patients in the RT completion and RT discontinuation groups, respectively (<i>P</i> = 0.294). <h3>Conclusion</h3> In this NCDB analysis, both clinical and socioeconomic characteristics, including age, gender, race, insurance status, income level, comorbidities, treatment type, and disease site, correlated with RT discontinuation rates. While toxicity of treatment was reported, patient decision was the most common reason for RT discontinuation. Table 1. Reasons for RT discontinuation (N = 212).